ORIGINAL RESEARCH ARTICLEpublished: 03 June 2014

doi: 10.3389/fnhum.2014.00360

Hand specific representations in language comprehensionClaire Moody-Triantis1, Gina F. Humphreys2 and Silvia P. Gennari1*

1 Department of Psychology, University of York, York, UK2 Neuroscience and Aphasia Research Unit, School of Psychological Sciences, University of Manchester, Manchester, UK

Edited by:

Analia Arevalo, East Bay Institute forResearch and Education, USA

Reviewed by:

Javier Gonzalez-Castillo, NationalInstitute of Mental Health, USALisa Aziz-zadeh, University ofSouthern California, USA

*Correspondence:

Silvia P. Gennari, Department ofPsychology, University of York, York,YO10 5DD, UKe-mail: silvia.gennari@york.ac.uk

Theories of embodied cognition argue that language comprehension involvessensory-motor re-enactments of the actions described. However, the degree of specificityof these re-enactments as well as the relationship between action and language remainsa matter of debate. Here we investigate these issues by examining how hand-specificinformation (left or right hand) is recruited in language comprehension and actionexecution. An fMRI study tested self-reported right-handed participants in two separatetasks that were designed to be as similar as possible to increase sensitivity of thecomparison across task: an action execution go/no-go task where participants performedright or left hand actions, and a language task where participants read sentencesdescribing the same left or right handed actions as in the execution task. We foundthat language-induced activity did not match the hand-specific patterns of activity foundfor action execution in primary somatosensory and motor cortex, but it overlapped withpre-motor and parietal regions associated with action planning. Within these pre-motorregions, both right hand actions and sentences elicited stronger activity than left handactions and sentences—a dominant hand effect. Importantly, both dorsal and ventralsections of the left pre-central gyrus were recruited by both tasks, suggesting differentaction features being recruited. These results suggest that (a) language comprehensionelicits motor representations that are hand-specific and akin to multimodal action plans,rather than full action re-enactments; and (b) language comprehension and actionexecution share schematic hand-specific representations that are richer for the dominanthand, and thus linked to previous motor experience.

Keywords: language comprehension, action execution, action representations, premotor cortex, left hand, right

hand, mirror neurons

INTRODUCTIONTheories of embodied cognition argue that language understand-ing implies partially simulating or re-enacting the actions beingdescribed and thus involves brain regions that are recruited inthe execution of those actions (Jeannerod, 2001; Glenberg andKaschak, 2002; Barsalou et al., 2003; Gallese and Lakoff, 2005;Barsalou, 2008). Indeed, it has been found that body part specificregions of the motor system are activated when reading lan-guage describing actions (Hauk et al., 2004; Buccino et al., 2005;Pulvermuller, 2005; Tettamanti et al., 2005) and they do so to aneffort specific degree (Moody and Gennari, 2010), suggesting thatlanguage recruits detailed action representations that would alsobe required for the execution of the same specific action.

However, the nature of the representations that are sharedbetween action and language remains unclear, in particular, theirlevel of specificity, i.e., to what extent do we re-enact the executionof an action described by language? Indeed, both primary motorand pre-motor regions have been associated with language com-prehension and these contrasting findings imply different levels ofspecificity in the representations elicited by language: if primarymotor regions are recruited during language comprehension,comprehenders can be thought to more closely re-enact the actiondescribed as if they were performing it, because these regionsare directly connected to the spinal cord and musculature (Dum

and Strick, 1996). Alternatively, if pre-motor and parietal regionsare recruited, comprehenders may activate more schematic actionplans that do not involve execution aspects per se, since pre-motorregions are typically associated with planning (Cisek et al., 2003).

The view that language may involve highly specific action rep-resentations is consistent with fMRI language studies that havereported the recruitment of primary motor regions (Hauk et al.,2004; Rüschemeyer et al., 2007; Kemmerer et al., 2008; Kemmererand Gonzalez-Castillo, 2010) and with TMS studies showing thatstimulation of primary motor cortex during language compre-hension modulates body-part specific motor evoked potentials(Oliveri et al., 2004; Buccino et al., 2005; Candidi et al., 2010). Incontrast, the view that language involves more schematic actionrepresentations is supported by many language studies show-ing the recruitment of planning-related pre-motor and parietalregions, rather than primary motor regions (Noppeney et al.,2005; Aziz-Zadeh et al., 2006; Moody and Gennari, 2010; Willemset al., 2010; Meteyard et al., 2012).

To shed light on this issue, we conducted an fMRI studydirectly comparing action execution and language comprehen-sion. The tasks were designed to be as similar as possible toincrease the sensitivity of the comparison. Every participantperformed an action execution and a language comprehensiontask. We focused on hand-specificity, i.e., whether the action is

Frontiers in Human Neuroscience www.frontiersin.org June 2014 | Volume 8 | Article 360 | 1

HUMAN NEUROSCIENCE

Moody-Triantis et al. Hand specificity in language

performed, or described as performed, with the left or the righthand. Importantly, the actions included in the execution task helda one-to-one correspondence with the content of the sentencesread in the language task. Thus, participants executed left andright hand button presses in the execution tasks and correspond-ingly read sentences describing left or right hand button presses inthe language task, albeit in different syntactic forms. In both tasks,participants were required to match a visual cue (e.g., L, R) refer-ring to a left or right hand action with the execution of the actionitself or the content of the sentence, thus keeping participantsfocused on the directionality of the stimuli.

This design has the potential of providing more homoge-neous activations and more precise and sensitive comparisonsacross conditions than previous studies. First, the linguistic stim-uli utilized refer to the same action, instead of classing togetherdifferent verbs (e.g., grasp, touch, give), which often have differentsenses and syntactic properties. Second, the linguistic meaningstargeted in the experiment had a one-to-one correspondencewith the actions executed in the execution task, unlike previousstudies comparing meaningless actions (e.g., finger movements)with semantically complex verbs (e.g., grasp) (Aziz-Zadeh et al.,2006). Finally, the execution task preceded the language taskto encourage imagery during the language task, thus increasingthe chances to detect potentially weak activity in primary motorregions.

Importantly, the focus on hand specificity provides sim-ple ways to distinguish between primary-motor and premotorregions, say, in comparison to body-part manipulations, becausethe activation patterns for left and right hand actions within pri-mary motor and pre-motor cortices is relatively well understood.Indeed, primary motor cortex has long been thought to play animportant role in the control of limbs on the contralateral sideof the body (Tanji et al., 1988; Dassonville et al., 1997; Aziz-Zadeh et al., 2002; Cisek et al., 2003). Thus, executing, observingor imagining a right-hand movement would recruit more neu-rons and stronger activity in the left primary motor cortex, andvise-versa. In contrast, activity in pre-motor regions responds toboth right and left hand actions both in cell recording and fMRIstudies (Tanji et al., 1988; Kermadi et al., 2000; Cisek et al., 2003;Hanakawa et al., 2006; Horenstein et al., 2009), although they mayrespond to different degrees (see below). This is due to the factthat the pre-motor cortex houses more schematic representationsresponsible for planning rather than executing actions, and thus,are less directly linked to the spinal cord (Rizzolatti and Luppino,2001).

Therefore, we predicted that if language comprehensioninvolves hand-specific representations, the pattern associatedwith either execution or planning of left and right handactions in primary motor or premotor areas should also beobserved in language comprehension. Specifically, if languagerecruits schematic planning representations only, then a sim-ilar pattern of activity across language comprehension andplanning should be found in pre-motor areas, but if linguis-tic representations are more detailed in execution content,language comprehension should match the execution-specificactivity pattern in primary motor regions, i.e., a contralateralpattern.

MATERIALS AND METHODSPARTICIPANTSEighteen participants were recruited for the experiment, allreported to be right-handed native English speakers with noknown neurological disorders, and to use the right hand in dailyand sport activities (14 female, 4 male; mean age 21, age range19–23 years).

MATERIALIn the execution tasks, visual letter cues were used to elicit but-ton presses that could include one or two fingers (e.g., LX, RX,LL, RR). In the language comprehension task, all sentences werewritten in the first person narrative (e.g., I am pressing. . . .) toencourage the activation of the participant’s own motor experi-ence during language comprehension. Each sentence describedleft/right hand button presses using either one or two fingers.In total 160 action sentences were presented. To encourage par-ticipants to process the sentence meaning and to maintain theirattention, the phrasing of the sentence was varied, for example,when describing one button press with the left hand participantscould read one of 4 different sentences (see Table 1). The lengthin characters of the sentences varied from 27 to 47 (mean length37.25), however to ensure that the sentences were all matchedacross conditions, the same structure was used in the left and rightconditions, with the words right and left varying accordingly.Therefore, psycholinguistic variables such as length and frequencyshould not influence the results.

TASK PROCEDURE AND DESIGNEthical approval for the study was obtained from Ethics com-mittee at the York Neuroimaging Centre, where the study wascarried out. Before the scanning session, participants were famil-iarized with the letter patterns to be used in the execution task.They practiced this task until they felt confident. They then prac-ticed the subsequent language task, which used the same cues butrequired different motor responses. All participants performedthe execution task before the language task.

Table 1 | Sentence stimuli.

Hand action Sentence

Right: two fingers I’m pressing both buttons with my right fingersI’m pushing two buttons on the rightI’m pushing two right buttonsOn the right, I’m pressing two buttons

Right: one finger I’m pressing the button with my right fingerI’m pushing one button on the rightI’m pushing one right buttonOn the right, I’m pressing one button

Left: two fingers I’m pressing both buttons with my left fingersI’m pushing two buttons on the leftI’m pushing two left buttonsOn the left, I’m pressing two buttons

Left: one finger I’m pressing the button with my left fingerI’m pushing one button on the leftI’m pushing one left buttonOn the left, I’m pressing one button

Frontiers in Human Neuroscience www.frontiersin.org June 2014 | Volume 8 | Article 360 | 2

Moody-Triantis et al. Hand specificity in language

Action execution taskA go-no-go task was used to elicit button presses. During theexperiment, participants held one button box in each hand rest-ing on their lap in a comfortable position. Each button box hadtwo buttons and participants were instructed to rest their indexand middle fingers on the buttons of the boxes during the experi-ment. Visual stimuli were projected through a mirror fixed to thehead coil. The go/no-go cues were pairs of letters in red upper-case 50 pt text. In total 200 action stimuli were presented, 160 gotrials and 40 no-go trials. The go trials instructed participants topress either one (RX, LX) or two buttons (RR, LL) using either theright or the left hand as quickly as possible (there were 40 trialsper cue). During practice, participant learned to match each let-ter of the visual cue onto each of the four buttons (and fingers), sothat RX indicated one button press with the right middle finger,RR indicated pressing both buttons simultaneously (middle andindex finger) and so on. The no-go trials instructed participantsto refrain from pressing a button (either XR or XL, i.e., an ini-tial X meant no response at all). Visual cues lasted for 500 ms andwere then replace by HH, which stayed on the screen until the nextcue. Cues from different conditions (left/right) using one or twofingers were intermixed in an event-related design following opti-mal stimulus order (the probability of each condition followingany other condition was constant) and random inter-trial timesobtained by a schedule optimizing algorithm (http://surfer.nmr.mgh.harvard.edu/optseq/). Therefore participants could not pre-dict the upcoming stimulus and had to plan each trial. Inter-trialinterval varied in duration from 2 to 26 s (average 5.8 s). The tasklasted 960 s in total.

Language comprehension taskParticipants remained in the scanner in the same positionand holding the same button boxes as in the previous task.Participants were presented with 160 sentences in white 30 pt text(on black background) each lasting 2000 ms and were asked toread the sentences for meaning. Table 1 exemplifies the differ-ent formats in which sentences were presented (10 cases of eachexample). After each sentence presentation, a sequence of 37 X’swere presented (which constitute the average character length ofall sentential stimuli) until next stimulus sentence appeared. Tokeep participants’ attention on the sentential content, 34 catch tri-als (also lasting 2000 ms) were also included in the design (21.25%of trials). As in action execution, an event-related design was usedwhere trial types were intermixed in such a way that the prob-ability of each trial type (sentence conditions plus catch trials)following any other type was constant, and therefore trial typescould not be predicted (the order of trials and inter-trial timeswere calculated with the same schedule optimizing algorithm asabove). Inter-trial intervals ranged from 2 to 26 s (average 4.96 s).Catch trials asked participants about the sentence content usingthe same cues that were used in the execution task, e.g., RR?Participants had to indicate whether the meaning of the previ-ously read sentence corresponded to the cue (meaning judgmenttask). To respond to this question, they had to use a left handbutton press (index finger for yes and middle finger for no).For example, participants may read I’m pushing two buttons onthe right, and after a few seconds (corresponding to the variable

inter-trial time), they may be presented with RR?, in which case,the correct answer is yes (a left index finger button press). In orderto perform well on this task participants had to read the sentencescarefully for their hand-specific action meaning, and therefore itensured that participants maintained their attention throughoutthe experiment.

DATA COLLECTION PARAMETERSA 3T GE Signa Exite MRI scanner was used to collect both high-resolution structural images and functional images. Functionalimages were obtained using a gradient-echo EPI sequence (TR2000 ms, TE 50 ms, flip angle 90◦, matrix 64 × 64, field of view24 cm) with 38 axial slices of thickness 3.0 mm. The resultingvoxel size was 3.75 × 3.75 × 3 cm. Note that our TE specificationis near those considered optimal for detecting signal in primarymotor cortex (Fera et al., 2004). Functional images excluded thecerebellum and in some participants inferior portions of thetemporal lobe. A T1 flair image was also obtained in order tofacilitate the registration between the high-resolution structuraland functional data.

DATA ANALYSISBoth first level and higher-level analyses were carried out forthe language and the action task separately using FEAT (FMRIExpert Analysis Tool) Version 5.91, part of FSL (FMRIB’sSoftware Library, www.fmrib.ox.ac.uk/fsl). We have followed thestandard order of processes built into the FSL FEAT analy-sis. Pre-processing steps included brain extraction, slice-timingcorrection, motion correction (Jenkinson et al., 2002), spatialsmoothing using a Gaussian kernel of FWHM 8 mm and high-pass temporal filtering (Gaussian-weighted least-squares straightline fitting, with sigma = 25.0 s). Time-series analysis was carriedout using FILM with local autocorrelation correction (Woolrichet al., 2001). A boxcar model of the timing of events was createdinvolving the onset and length of each stimulus event, which wasthen convolved with a hemodynamic response (gamma) function.For both action and language data the events were modeled atthe onset of the stimulus presentation with action trials lasting500 ms and language trials lasting 2000 ms. For the language task,the catch trials were modeled separately to partial out the par-ticipant’s motor responses but were excluded from any statisticalaverage or comparison of the language data. No-go trials in theexecution task were also modeled out and not analyzed further.

Several contrasts were run at the individual level between thedifferent conditions in the execution and the language task. Forboth the execution and the language data, all actions or sentencestogether (irrespective of hand) were compared to rest to iden-tify all action or all language regions, and right and left handactions or sentences were also compared against one another tofind those areas that were significantly more involved in perform-ing or reading about left or right hand actions (R > L, L > R).Individual level analyses were then entered into high-level mixed-effect modeling built into FSL, taking into account both varianceand parameter estimates from individual-level results. All higher-level analyses reported below were carried out using FLAME(FMRIB’s Local Analysis of Mixed Effects (Woolrich et al., 2004)within the right or left hemisphere to increase statistical power.

Frontiers in Human Neuroscience www.frontiersin.org June 2014 | Volume 8 | Article 360 | 3

Moody-Triantis et al. Hand specificity in language

Z (Gaussuanised T/F) statistic images were thresholded using aGaussian Random Field-theory (GRF)-based maximum heightwith a (corrected) significance threshold of p = 0.05 (Worsleyet al., 1992). For convenience, we will refer to this correctionmethod, GRF-based correction.

Region of interest analyses in hand- and execution-specific regionsTo evaluate whether language activity within primary motorregions showed the same pattern as that of action execu-tion, we used execution-specific activity to identify regions forfurther analyses of hand-specific language activity. To isolatehand-specific execution regions that would not include commonplanning regions, we used execution activity resulting from con-trasting left-hand and right-hand actions, i.e., the contrasts R > Land L > R, obtained with GRF-based correction at p = 0.05.Subtracting left from right and right from left action perfor-mance should cancel out any general planning activity that isshared across hands, thus identifying execution specific activity,which should show the typical contralateral pattern. Indeed, sim-ply comparing left-hand or right-hand execution relative to restmay still include regions that are common to both hands, and thuslikely to reflect common planning regions, because these gen-eral contrasts only identify voxels active for one hand irrespectiveof the other hand. These contrasts yielded as expected, the con-tralateral pattern shown in in Figure 1 in the blue-to-cyan andred-to-yellow scales. Within these hand- and execution-specificcontralateral ROI masks, we then ran a high-level analysis (GRF-based correction, p = 0.05) for the language data irrespective ofhand, i.e., the contrast all sentences vs. rest, to establish whetherlanguage comprehension activated these hand-specific executionregions. This yielded significant language activity (irrespective ofhand) shown in green in Figure 1. The average percent signalchange within the significant cluster resulting from this high-levelanalysis was then extracted for each participant using FSL tools.T-tests (with subjects as random factor) were then used to deter-mine whether there was any difference between left-hand andright-hand sentences.

Region of interest analyses in non-hand-specific regionsTo isolate regions that were sensitive to all hand actions irre-spective of hand and thus were likely to include activations in

FIGURE 1 | Results from the action execution task showing the

contralateral pattern of activation specifically responding to left hand

actions (in blue, left hand > right hand contrast) and right hand

actions (in red, right hand > left hand) (whole brain GRF-based

correction, p = 0.05). Significant language comprehension activityresponding to all sentence types within each execution region is shown ingreen.

planning regions, we contrasted all actions relative to rest (GRF-based correction, p = 0.05). The corresponding contrast was alsoconducted in the language task to identify all regions involvedin language comprehension irrespective of hand (GRF-based cor-rection, p = 0.05). By multiplying these execution and languagecomprehension results, we localized several clusters that were sig-nificantly active in both tasks, and thus indicated overlappingregions across tasks. This is equivalent to a conjunction analy-ses as previously referred to in the literature (Nichols et al., 2005).These overlapping clusters thus acted as functional localizers forthe regions targeted for further analysis of more specific con-trasts (Poldrack, 2007). In particular, to establish whether therewere hand-specific activations within these overlapping regions,we extracted the percent signal changes for each hand relative torest for each participant in each of the main overlapping clustersshown in Figure 2. These values were then analyzed with pairedt-tests (with subjects as random factor) to examine whether eitherin action planning or in language comprehension, there wasstronger activity for a specific hand, and more generally, to exam-ine whether a similar pattern of activity was shown for planningand language, as hypothesized.

RESULTSBEHAVIORAL DATAExecution taskThe time taken to perform the instructed action and the numberof errors made were measured.

Reaction times. Trials containing errors or responses longer than3 standard deviations from the mean were excluded from the reac-tion times analyses. These exclusions constituted about 3.40%

FIGURE 2 | Action execution activity (in blue) and language

comprehension activity (in red) in response to all actions and all

sentence stimuli compared to rest (whole brain GRF-based correction,

p = 0.05). The regions in which language and execution activity overlapped(conjunction) are shown in green and are labeled as dorsal pre-motor (dPM),ventral pre-motor (vPM) and parietal lobe (PL).

Frontiers in Human Neuroscience www.frontiersin.org June 2014 | Volume 8 | Article 360 | 4

Moody-Triantis et al. Hand specificity in language

of the total data. We found that participants responded fasterwith the right hand (mean = 615.8 ms) than the left hand(mean = 630.2 ms) [t(18) = 2.77, p = 0.01], thus providing sup-porting evidence that our participants were indeed right-handed.

Accuracy. A response was classed as an error if participant eitherfailed to make a response or responded using the wrong hand.On average participants made an error on 3.06% of action tri-als, although there was not reliable difference between left andright hand actions (Wilcoxon Signed-Rank test: z = −0.637,p > 0.05). The numbers of errors were also calculated on no-go trials, with errors being defined as those no-go trials wherean action was incorrectly performed. On average, errors on no-go trials were relatively low and were made 2.5% of the time.Furthermore, almost all errors (94%) were consistent with thedirectional letter in the cue (i.e., if the cue was XR the right buttonwas most likely to be erroneously pressed, and vise-versa).

Language taskDue to experimenter error, no responses were recorded from oneparticipant. For the remaining 17 participants, on average partici-pants responded correctly on 90.7% of the question trials and themean reaction time for the responses was 2605 ms, as measuredfrom the presentation of the cue (e.g., RR?).

OVERALL FUNCTIONAL ACTIVATIONS FOR ACTION EXECUTION ANDLANGUAGE COMPREHENSIONAction representations in hand- and execution-specific regionsAs anticipated from previous research, hand-specific action exe-cution (left > right and right > left) elicited stronger responsesin the contralateral hemisphere (GRF-based correction, p = 0.05)(Figure 1). The strongest activity was centered around the postcentral gyrus and extended into the central sulcus and pre-centralgyrus (left hemisphere peak: −40, −26, 54; right hemispherepeak: 42, 4–30, 58). The corresponding corrected analysis for thelanguage comprehension data contrasting one hand relative to theother however did not elicit any significant response. To makesure that stringent correction level did not miss hand-specificlanguage activity, we conducted further ROI analyses within thecontralateral execution clusters, as described in Region of InterestAnalyses in Hand- and Execution-specific Regions and reportedbelow.

Actions representations in non-hand-specific (planning) regionsThe contrast of all actions relative to rest (GRF-based correction,p = 0.05) revealed several brain regions that were commonly acti-vated by the execution task irrespective of hand. These includedpremotor and parietal regions, as well as other regions. Peakactivations for the left-hemisphere are listed in Table 2, and theoverall pattern of execution activity is shown in the blue-to-cyanscale in Figure 2. The contrast of all sentences relative to rest alsorevealed several brain regions that included parietal, pre-motor,posterior temporal and inferior frontal regions (GRF-based cor-rection, p = 0.05). Peak activations for the left-hemisphere arelisted in Table 2, and the overall language activity is shown in theyellow-to-red scale in Figure 2. The multiplication of the activityelicited by each of these tasks indicated regions that were signifi-cantly activated for both action execution/planning and language

Table 2 | Peak activations for each task and center of gravity for

overlapping regions.

Tasks Anatomical label MNI x,y,z z Voxels in

cluster

Action execution Post central gyrus −4, −50, 74 4.87 60

Precentral gyrus −32, −10, 52 6.51 415

−60, 4, 14 5.58 59

−54, 6, 28 4.57 57

−48, −4, 42 5.45 15

Cingulate gyrus/SMA −8, 2, 42 5.98 391

SMA 2, −6, 70 14

Supramarginal gyrus −46, −36, 42 6.64 1047

Superior frontal gyrus −20, −6, 72 5.33 22

Opercular cortex −48, 0, 0 5.3 29

Lateral occipital cortex −40, −74, 0 5.23 97

−22, −70, 32 4.68 14

Languagecomprehension

Pre−central gyrus −44, 0, 42 6.54 926

Inferior frontal gyrus −56, 14, 18 5.5 110

Middle temporal gyrus −54, −48, 6 6 191

SMA −2, −2, 70 5.73 207

Precuneous 0, −68, 60 5.89 566

Superior parietal lobule −32, −60, 42 6.22 166

Fusiform cortex −44, −44, −20 6.14 110

Lateral occipital cortex −24, −72, 30 5.28 28

Execution +language areas(center ofgravity)

Precentral gyrus/middlefrontal gyrus (dorsalpremotor—dPM)

−34, −7, 55 419

Precentral Gyrus (ventralpre-motor—vPM)

−50, 0, 33 351

Parietal lobe −34, −52, 40 410

Supplementary motorarea

−1, 3, 52 324

Coordinates are given for the left hemisphere, which were analogous to those

in the right hemisphere. Cluster sizes are given for the GRF-based corrected

images at a threshold of z = 4.5.

(conjunction), as shown in green in Figure 2. These commonactivations suggest that common neural representations wererecruited for both execution/planning and language comprehen-sion. The overlapping regions were located in the middle frontalgyrus/dorsal pre-central gyrus, superior parietal lobule/angulargyrus, and ventral pre-central gyrus and were larger in the leftthan the right hemisphere. Because these regions were associatedwith more than one anatomical label according to the Harvard-Oxford Cortical Structural Atlas, henceforth we refer to them asdorsal or ventral pre-motor regions (dPM, vPM) and parietal loberegions (PL). The centers of gravity of these regions are listed inTable 2.

REGION OF INTERESTSHand- and execution-specific regionsHand specific language activity was assessed in two steps(see section Region of Interest Analyses in Hand- and

Frontiers in Human Neuroscience www.frontiersin.org June 2014 | Volume 8 | Article 360 | 5

Moody-Triantis et al. Hand specificity in language

Execution-specific Regions) because direct contrast between left-and right-hand sentences did not show any significant voxel ina high-level analysis masked by the hand- and execution-specificROIs of Figure 1. We first conducted a high-level analysis withinhand specific execution ROIs to detect any language activity irre-spective of hand (all sentences vs. rest). This analysis revealedsignificant clusters shown in green in Figure 1. The clusters werelocated in the superior portion of the pre-central gyrus (lefthemisphere peak: −32, −10, 64). Within these clusters, we thenevaluated hand-specific activity by extracting the percent signalchange for left and right hand sentences vs. rest for each partici-pant and for each of the left and right hemisphere clusters. T-testscomparing left vs. right hand sentence activity within these clus-ters revealed no significant difference (p > 0.4). The hand specificpattern of data as seen in action performance is therefore not seenwhen comprehending hand specific action language within theseexecution areas.

Non-hand specific (planning) regionsTo examine whether a similar pattern of activity was shownfor planning and language within the regions that were sig-nificantly activated in both tasks, as hypothesized, for each ofthe identified common regions of activation for the languageand execution tasks (see above and Figure 2), we contrastedright and left hand actions or sentences for each of the hemi-spheres. The overall pattern of results is summarized in Figure 3.For all the common clusters of activation in the left hemi-sphere, we found a parallel pattern of activation across actionexecution and language comprehension. As shown in Figure 3,right-hand actions or sentences elicited stronger activity than left-hand actions or sentences [dPM—language activity: t(17) = 2.71,p < 0.02; dPM—execution activity: t(17) = 5.98, p = 0.0001;PL—language activity: t(17) = 3.42, p < 0.003; PL—execution

activity: t(17) = 2.46, p < 0.03; vPM—language activity: t(17) =2.53, p < 0.01; vPM—execution activity: t(17) = 2.53, p < 0.02].For the common clusters of activation in the right hemisphere,the pattern of results was numerically similar to that in theleft hemisphere, with right-hand actions or sentences also elic-iting a stronger response than left hand actions or sentences.However, only the vPM cluster showed statistically significantresults for execution and language [language activity: t(17) =3.61, p < 0.002; execution activity: t(17) = 2.96, p < 0.009], withall other right-hemisphere regions not reaching significance(p > 0.05). Note that these results, and particularly those inpre-motor regions, could not be due to eye-movements duringreading, which we could not control for: First, left vs. right sen-tences were identical except for one word, and thus are likely toelicit similar eye-movements. Therefore, the differences in hand-specific activity cannot be due to more or less eye-movement inone condition relative to other. Second, the coordinate range typ-ically associated with the frontal-eye field (Paus, 1996; Swallowet al., 2003) do not correspond to those reported here, consis-tent with the fact that this region is anterior to the hand area.Finally, the execution task, with which language activity over-lapped, only involved central fixation, and therefore, cannot bedue to eye-movements.

Overall, these results suggest that hand-specific effects arefound in regions of common activity for action planning and lan-guage comprehension in left pre-motor and parietal regions andright pre-motor regions. Because these regions were active forthe execution of either hand action and were not located in pri-mary motor regions, they reflect more schematic representationsassociated with planning, rather than muscle control. Therefore,action execution/planning and language comprehension appearto recruit some aspects of these more schematic representations.Interestingly, both language comprehension and action execution

FIGURE 3 | Percent signal change for right or left hand actions and right or left hand sentences within regions of overlap between action execution

and language comprehension (see Figure 2). All comparisons are significant at p < 0.02. Error bars represent standard error.

Frontiers in Human Neuroscience www.frontiersin.org June 2014 | Volume 8 | Article 360 | 6

Moody-Triantis et al. Hand specificity in language

show hand specific effects characterized by stronger responsesfor the right hand than for those of the left hand, suggestinga dominant hand effect, since our participants reported to beright-handed. We will discuss this specific effect below.

DISCUSSIONThis study aimed to investigate the nature of the representationsthat are recruited by hand-specific information during languagecomprehension, and to assess the extent to which we simulatethe actions that we read about by comparing language activityto motor-related activity elicited by similar tasks. Participantswere asked to perform left and right hand button presses andread sentences that described the same left and right hand but-ton presses. Hand-specific activity for the language task was thenassessed within the primary motor hand-specific contralateralregions where execution and language activity overlapped. Wepredicted that if we nearly accurately re-enact the actions weread about, hand-specific contralateral activity should occur inexecution-specific areas such as primary motor cortex for lan-guage comprehension in the same way that it does in actionexecution. This prediction was not supported. Although therewas some significant language activity in the superior portion ofthe pre-central gyrus, a contralateral pattern of activity for hand-specific actions or any sensitivity to hand-specificity was not seenin language as it was in action execution (section Hand- andExecution-specific Regions). This suggests that language com-prehension does not show sensitivity to hand-specificity withinthese execution areas, and therefore that the hand-specific infor-mation that was required for language comprehension was notrepresented within execution areas.

We also predicted that if hand-specific information is repre-sented in a more schematic and general way during languagecomprehension, then those areas that are responsible for actionplanning (including the premotor and parietal cortex) would dis-play equivalent activation patterns in the action execution andlanguage task for left and right hand actions. This predictionwas assessed in those regions of the premotor and parietal cortexthat were activated during action execution and language compre-hension irrespective of hand, i.e., these regions were significantlyactive for both left and right hand action or sentences [sec-tion Actions Representations in Non-hand-specific (Planning)Regions], but we further examined whether there was any hand-specific differences in the amplitude of this activation [sectionNon-hand Specific (Planning) Regions]. We found that there wasmore activity for right-hand actions and right-hand sentencesthan left ones in most of the pre-motor and parietal regions exam-ined within the left hemisphere (a dominant hand effect) as wellas in pre-motor areas of the right hemisphere. This indicates asimilar pattern of activation across language comprehension andaction execution/planning in pre-motor regions, as predicted.Together these results provide support for embodied cognitionand suggest that language recruits detailed hand-specific actionrepresentations that are nevertheless one-step removed fromre-enacting the execution of the action itself. In other words,language comprehension does not fully activate all action com-ponents that are required for the performance of that action.Instead, only more schematic action representations that are

stored in areas responsible for action planning are recruited forlanguage.

The dominant hand effect, i.e., that right-hand actions or sen-tences elicited more activity than left-hand ones in pre-motorregions, is consistent with previous studies suggesting that motorrepresentations in language comprehension and action observa-tion are modulated by motor experience (e.g., Buccino et al.,2004; Calvo-Merino et al., 2005; Beilock et al., 2008). Indeed,language studies have shown that right handers and left handersactivate pre-motor cortex to a different degree in different hemi-spheres (Willems et al., 2010), and activity in pre-motor regionsdescribing hockey actions correlates with different degrees ofhockey experience in the dominant hemisphere (Beilock et al.,2008). In action observation studies, more activity is also seen inpre-motor areas for observing human compared to non-humanactions (Buccino et al., 2004), biomechanically performableactions compared to non-performable actions (Costantini et al.,2005; Candidi et al., 2008) or those actions that a participant isexpert, rather than inexperienced in performing (Calvo-Merinoet al., 2005; Haslinger et al., 2005; Cross et al., 2006; Kiefer et al.,2007; Beilock et al., 2008). This suggests that increased expe-rience results in the establishment of a more elaborate actionrepresentation leading to stronger activations in action execution,observation, and language comprehension.

Our results are consistent with much of the literature onpre-motor cortex showing that unlike primary motor regions,ventral and dorsal premotor regions play a variety of a cogni-tive functions supporting not only action planning, e.g., via theformation of visuo-motor associations, but also perceptual anal-ysis, serial prediction and attentional functions (Johnson et al.,1996; Boussaoud, 2001; Picard and Strick, 2001; Simon et al.,2002; Schubotz and von Cramon, 2003; Cisek and Kalaska, 2004;Chouinard et al., 2005). In particular, this research has proposedfunctional differentiations between dorsal and ventral portionsof the premotor cortex (e.g., Schubotz and von Cramon, 2003).In this respect, our results suggest common representations forexecution/planning and language comprehension in these twopremotor regions, as we found a more dorsal pre-motor clusterin the left hemisphere and another cluster more ventral and bilat-eral (Figure 2). Although both left hemisphere clusters are locatedin the proximity of previously reported hand-related motor andlanguage activity, which indeed have been reported to be locatedeither more dorsally or ventrally (see summary of coordinates inKemmerer and Gonzalez-Castillo, 2010), the fact that two dis-tinct clusters were fund here suggests different roles for theseregions. More dorsal aspects of the pre-motor cortex are impli-cated in spatial attention and specifically, the use of current orexpected sensory features of the environment relative to the body(Boussaoud, 2001; Schubotz and von Cramon, 2003), which isconsistent with the attention to directionality required in bothour tasks relative to the body. Therefore, it is possible that dif-ferent aspects of the action representation are distributed acrossthe pre-motor cortex, one cluster linked to spatial features andanother to motor plans or schemas.

More importantly for the purpose of our study, our resultshave implications for theories of embodied cognition as appliedto language. Although we cannot exclude the possibility that

Frontiers in Human Neuroscience www.frontiersin.org June 2014 | Volume 8 | Article 360 | 7

Moody-Triantis et al. Hand specificity in language

other more sensitive methods or more targeted designs may reveallanguage sensitivity in primary motor regions, the same exper-imental conditions that elicited significant effects in pre-motorregions were not sufficient to detect hand-sensitive activity in pri-mary motor regions. Thus, the comprehension of hand actionsentences does not seem to involve action representations thatare specifically recruited for left or right hand executions in con-tralateral hemispheres, even when imagery was encouraged by theorder and similarity of the execution and language tasks. This sug-gests that those regions of primary motor cortex directly linked tothe spinal cord are not activated by language and language-elicitedimagery in similar conditions to those that activate pre-motorregions. This contrasts with previous fMRI and TMS reports,which may have been tapping into planning components and didnot distinguish between effector-specific plans and executions. InTMS studies in particular, it is very likely that stimulation of pri-mary motor cortex will stimulate pre-motor cortex too, due tostrong interconnections between the two (Chouinard et al., 2003).Therefore, language does not appear to elicit simulations of theaction described as if we were performing the action, but rather asif we had the intention or idea of performing the action.

Nevertheless, we do find stronger activity for the dominantright hand bilaterally in the pre-central gyrus, and in other leftpre-motor and parietal regions. According to previous findings,this suggests that action plans or schemas in these regions activatericher representations for the dominant hand, and in this respect,they are hand-specific representations, i.e., they include informa-tion as to whether the action would be executed with the left orthe right hand. This is particularly revealing because previous lan-guage studies have suggested that hand-action representations arebody specific, i.e., right and left handers display opposing activa-tion patterns across the hemispheres in premotor regions, withright handers showing more activity in the right hemisphere thanthe left hemisphere and vice-versa (Willems et al., 2010). Here, wego a step further and show that these pre-motor representationsare not only body-specific but also hand-specific. Even more, ifthe rich experience associated with the dominant hand is indeedresponsible for stronger activity, our results suggests that handdominance is not only represented on the dominant hemispherebut also bilaterally in the pre-central gyrus, suggesting sharedfunctions across the hemispheres.

These observations are consistent with the fact that mirrorneurons have primarily been reported in pre-motor regions,rather than primary motor ones, and are considered multimodal,often integrating visual, somatosensory and motor information(e.g., Rizzolatti et al., 2002; Gallese and Lakoff, 2005). It is thusconceivable that language may also activate them, particularlyin a task where attention to hand effector and directionality isrequired. However, these partial re-enactments only support orcontribute to language comprehension, as other regions were alsorecruited for language comprehension but not action execution,most notably, the left inferior frontal gyrus and the posteriortemporal lobe (see Figure 2). These two regions have been consis-tently implicated in many lesion and imaging studies of languageprocessing (Jefferies and Lambon Ralph, 2006; Tyler and Marslen-Wilson, 2008; Humphreys and Gennari, 2014), suggesting thattheir role is critical to language comprehension. Therefore, our

study demonstrates those aspects of the language network whereaction representations are shared with action planning.

Nevertheless, the cognitive role of mirror-like activity in thebrain still remains to be fully understood. Recent findings suggestthat mirror-like responses can also be found in primary motorcortex, and that canonical mirror responses can also be foundin the hippocampus, SMA and medial frontal regions (Tkachet al., 2007; Lepage et al., 2008; Mukamel et al., 2010). These sameregions also display cells with opposite pattern of excitation andinhibition to those observed during action execution or obser-vation, suggesting a role for both integration and differentiationof representations across the brain. Complex activity patternsof neural assemblies across the brain have already been stud-ied in detail by researchers interested in the control of behavior,for which attention and working memory (the need to main-tain a goal in memory through complex sequences of actions) arekey cognitive processes (e.g., Fuster, 2001). This sort of systemicapproach, where temporally integrated activity patterns are inves-tigated across a large network, is likely to provide critical clues forunderstanding emergent cognitive processes.

CONCLUSIONThe present results suggest that within the constraints andassumptions of fMRI research, we don’t appear to re-enact theactions that we read about in all the same brain areas thatare required for action execution. Only very particular actionrepresentations are recruited by language—those involved inmore abstract stages of action planning in pre-motor cortex.Nevertheless, the representations that are stored in these plan-ning regions are highly specific in that they contain hand-specificinformation. This is therefore consistent with embodied theo-ries of language proposing that language understanding involvesthe partial re-enactment of the action described, including hand-specific representations, but we do not accurately re-enact theaction as such throughout the motor system. Language under-standing is therefore somewhat removed from action executionas it relies upon higher-level cognitive regions.

REFERENCESAziz-Zadeh, L., Maeda, F., Zaidel, E., Mazziotta, J., and Iacoboni, M.

(2002). Lateralization in motor facilitation during action observation:a TMS study. Exp. Brain Res. 144, 127–131. doi: 10.1007/s00221-002-1037-5

Aziz-Zadeh, L., Wilson, S. M., Rizzolatti, G., and Iacoboni, M. (2006). Congruentembodied representations for visually presented actions and linguistic phrasesdescribing actions. Curr. Biol. 16, 1818–1823. doi: 10.1016/j.cub.2006.07.060

Barsalou, L. W. (2008). Grounded cognition. Annu. Rev. Psychol. 59, 617–645. doi:10.1146/annurev.psych.59.103006.093639

Barsalou, L. W., Simmons, W. K., Barbey, A. K., and Wilson, C. D. (2003).Grounding conceptual knowledge in modality-specific systems. Trends Cogn.Sci. 7, 84–91. doi: 10.1016/S1364-6613%2802%2900029-3

Beilock, S. L., Lyons, I. M., Mattarella-Micke, A., Nusbaum, H. C., and Small, S.L. (2008). Sports experience changes the neural processing of action language.Proc. Natl. Acad. Sci. U.S.A. 105 , 13269–13273. doi: 10.1073/pnas.0803424105

Boussaoud, D. (2001). Attention versus intention in the primate premotor cortex.Neuroimage 14(1 Pt 2), S40–S45. doi: 10.1006/nimg.2001.0816

Buccino, G., Lui, F., Canessa, N., Patteri, I., Lagravinese, G., Benuzzi, F., et al.(2004). Neural circuits involved in the recognition of actions performedby nonconspecifics: an FMRI study. J. Cogn. Neurosci. 16, 114–126. doi:10.1162/089892904322755601

Frontiers in Human Neuroscience www.frontiersin.org June 2014 | Volume 8 | Article 360 | 8

Moody-Triantis et al. Hand specificity in language

Buccino, G., Riggio, L., Melli, G., Binkofski, F., Gallese, V., and Rizzolatti, G. (2005).Listening to action-related sentences modulates the activity of the motor sys-tem: a combined TMS and behavioral study. Cogn. Brain Res. 24, 355–363. doi:10.1016/j.cogbrainres.2005.02.020

Calvo-Merino, B., Glaser, D. E., Grèzes, J., Passingham, R. E., and Haggard, P.(2005). Action observation and acquired motor skills: an fMRI study withexpert dancers. Cereb. Cortex 15, 1243–1249. doi: 10.1093/cercor/bhi007

Candidi, M., Leone-Fernandez, B., Barber, H. a, Carreiras, M., and Aglioti, S. M.(2010). Hands on the future: facilitation of cortico-spinal hand-representationwhen reading the future tense of hand-related action verbs. Eur. J. Neurosci. 32,677–683. doi: 10.1111/j.1460-9568.2010.07305.x

Candidi, M., Urgesi, C., Ionta, S., and Aglioti, S. M. (2008). Virtual lesion of ven-tral premotor cortex impairs visual perception of biomechanically possible butnot impossible actions. Soc. Neurosci. 3, 388–400. doi: 10.1080/17470910701676269

Chouinard, P. A., Leonard, G., and Paus, T. (2005). Role of the primary motorand dorsal premotor cortices in the anticipation of forces during object lifting.J. Neurosci. 25, 2277–2284. doi: 10.1523/JNEUROSCI.4649-04.2005

Chouinard, P. A., Van Der Werf, Y. D., Leonard, G., and Paus, T. (2003). Modulatingneural networks with transcranial magnetic stimulation applied over the dor-sal premotor and primary motor cortices. J. Neurophysiol. 90, 1071–1083. doi:10.1152/jn.01105.2002

Cisek, P., Crammond, D. J., and Kalaska, J. F. (2003). Neural activity in primarymotor and dorsal premotor cortex in reaching tasks with the contralateral versusipsilateral arm. J. Neurophysiol. 89, 922–942. doi: 10.1152/jn.00607.2002

Cisek, P., and Kalaska, J. F. (2004). Neural correlates of mental rehearsal in dorsalpremotor cortex. Nature 431, 993–996. doi: 10.1038/nature03005

Costantini, M., Galati, G., Ferretti, A., Caulo, M., Tartaro, A., and Romani,G. L. (2005). Neural systems underlying observation of humanly impossi-ble movements: an FMRI study. Cereb. Cortex 15, 1761. doi: 10.1093/cercor/bhi053

Cross, E. S., Hamilton, A. F., De, C., and Grafton, S. (2006). Building a motor sim-ulation de novo: observation of dance by dancers. Neuroimage 31, 1257–1267.doi: 10.1016/j.neuroimage.2006.01.033

Dassonville, P., Zhu, X.-H., Ugurbil, K., Kim, S.-G., and Ashe, J. (1997). Functionalactivation in motor cortex reflects the direction and the degree of?handedness.Proc. Natl. Acad. Sci. U.S.A. 94, 14015–14018. doi: 10.1073/pnas.94.25.14015

Dum, R. P., and Strick, P. L. (1996). Spinal cord terminations of the medial wallmotor areas in macaque monkeys. J. Neurosci. 16, 6513–6525.

Fera, F., Yongbi, M. N., van Gelderen, P., Frank, J. A., Mattay, V. S., and Duyn, J. H.(2004). EPI-BOLD fMRI of human motor cortex at 1.5 T and 3.0 T: sensitivitydependence on echo time and acquisition bandwidth. J. Magn. Reson. Imaging19, 19–26. doi: 10.1002/jmri.10440

Fuster, J. M. (2001). The prefrontal cortex–an update: time is of the essence. Neuron30, 319–333. doi: 10.1016/S0896-6273(01)00285-9

Gallese, V., and Lakoff, G. (2005). The Brain’s concepts: the role of the Sensory-motor system in conceptual knowledge. Cogn. Neuropsychol. 22, 455–479. doi:10.1080/02643290442000310

Glenberg, A. M., and Kaschak, M. P. (2002). Grounding language in action.Psychon. Bull. Rev. 9, 558–565. doi: 10.3758/BF03196313

Hanakawa, T., Honda, M., Zito, G., Dimyan, M., and Hallett, M. (2006). Brainactivity during visuomotor behavior triggered by arbitrary and spatially con-strained cues: an fMRI study in humans. Exp. Brain Res. 172, 275–282. doi:10.1007/s00221-005-0336-z

Haslinger, B., Erhard, P., Altenmüller, E., Schroeder, U., Boecker, H., and Ceballos-Baumann, A. O. (2005). Transmodal sensorimotor networks during actionobservation in professional pianists. J. Cogn. Neurosci. 17, 282–293. doi:10.1162/0898929053124893

Hauk, O., Johnsrude, I., and Pulvermuller, F. (2004). Somatotopic representation ofaction words in human motor and premotor cortex. Neuron 41, 301–307. doi:10.1016/S0896-6273(03)00838-9

Horenstein, C., Lowe, M. J., Koenig, K. A., and Phillips, M. D. (2009). Comparisonof unilateral and bilateral complex finger tapping-related activation in pre-motor and primary motor cortex. Hum. Brain Mapp. 30, 1397–1412. doi:10.1002/hbm.20610

Humphreys, G. F., and Gennari, S. P. (2014). Competitive mechanisms in sen-tence processing: common and distinct production and reading comprehen-sion networks linked to the prefrontal cortex. Neuroimage 84, 354–366. doi:10.1016/j.neuroimage.2013.08.059

Jeannerod, M. (2001). Neural simulation of action: a unifying mecha-nism for motor cognition. Neuroimage 14(1 Pt 2), S103–S109. doi:10.1006/nimg.2001.0832

Jefferies, E., and Lambon Ralph, M. A. (2006). Semantic impairment in strokeaphasia versus semantic dementia: a case-series comparison. Brain 129,2132–2147. doi: 10.1093/brain/awl153

Jenkinson, M., Bannister. P., Brady, M., and Smith, S. (2002). Improved optimi-sation for the robust and accurate linear registration and motion correction ofbrain images. Neuroimage 17, 825–841. doi: 10.1006/nimg.2002.1132

Johnson, P. B., Ferraina, S., Bianchi, L., and Caminiti, R. (1996). Cortical net-works for visual reaching: physiological and anatomical organization of frontaland parietal lobe arm regions. Cereb. Cortex 6, 102–119. doi: 10.1093/cercor/6.2.102

Kemmerer, D., Castillo, J. G., Talavage, T., Patterson, S., and Wiley, C. (2008).Neuroanatomical distribution of five semantic components of verbs: evidencefrom fMRI. Brain Lang. 107, 16–43. doi: 10.1016/j.bandl.2007.09.003

Kemmerer, D., and Gonzalez-Castillo, J. (2010). The two-level theory of verb mean-ing: an approach to integrating the semantics of action with the mirror neuronsystem. Brain Lang. 112, 54–76. doi: 10.1016/j.bandl.2008.09.010

Kermadi, I., Liu, Y., and Rouiller, E. M. (2000). Do bimanual motor actions involvethe dorsal premotor (PMd), cingulate (CMA) and posterior parietal (PPC)cortices? Comparison with primary and supplementary motor cortical areas.Somatosens. Motor Res., 17, 255–271. doi: 10.1080/08990220050117619

Kiefer, M., Eun-Jin, S., Liebich, S., Hauk, O., and Tanaka, J. (2007). Experience-dependent plasticity of conceptual representations in human sensory-motorareas. J. Cogn. Neurosci. 19, 525–542. doi: 10.1162/jocn.2007.19.3.525

Lepage, J.-F., Lortie, M., and Champoux, F. (2008). Action-coding neurons in pri-mary motor cortex: making sense of M1 activity during action perception.J. Neurosci. 28, 1995–1996. doi: 10.1523/JNEUROSCI.5422-07.2008

Meteyard, L., Cuadrado, S. R., Bahrami, B., and Vigliocco, G. (2012). Coming ofage: a review of embodiment and the neuroscience of semantics. Cortex 48,788–804. doi: 10.1016/j.cortex.2010.11.002

Moody, C. L., and Gennari, S. P. (2010). Effects of implied physical effortin sensory-motor and pre-frontal cortex during language comprehension.Neuroimage 49, 782–793. doi: 10.1016/j.neuroimage.2009.07.065

Mukamel, R., Ekstrom, A. D., Kaplan, J., Iacoboni, M., and Fried, I. (2010). Single-neuron responses in humans during execution and observation of actions. Curr.Biol. 20, 750–756. doi: 10.1016/j.cub.2010.02.045

Nichols, T., Brett, M., Andersson, J., Wager, T., and Poline, J.-B. (2005). Valid con-junction inference with the minimum statistic. Neuroimage 25, 653–660. doi:10.1016/j.neuroimage.2004.12.005

Noppeney, U., Josephs, O., Kiebel, S., Friston, K. J., and Price, C. J. (2005). Actionselectivity in parietal and temporal cortex. Cogn. Brain Res. 25, 641–649. doi:10.1016/j.cogbrainres.2005.08.017

Oliveri, M., Finocchiaro, C., Shapiro, K., Gangitano, M., Caramazza, A., andPascual-Leone, A. (2004). All talk and no action: a transcranial magnetic stimu-lation study of motor cortex activation during action word production. J. Cogn.Neurosci. 16, 374–381. doi: 10.1162/089892904322926719

Paus, T. (1996). Location and function of the human frontal eye-field: a selectivereview. Neuropsychologia 34, 475–483. doi: 10.1016/0028-3932(95)00134-4

Picard, N., and Strick, P. L. (2001). Imaging the premotor areas. Curr. Opin.Neurobiol. 11, 663–672. doi: 10.1016/S0959-4388(01)00266-5

Poldrack, R. A. (2007). Region of interest analysis for fMRI. Soc. Cogn. Affect.Neurosci. 2, 67–70. doi: 10.1093/scan/nsm006

Pulvermuller, F. (2005). Brain mechanisms linking language and action. Nat. Rev.Neurosci. 6, 576–582. doi: 10.1038/nrn1706

Rizzolatti, G., Fogassi, L., and Gallese, V. (2002). Motor and cognitive func-tions of the ventral premotor cortex. Curr. Opin. Neurobiol. 12, 149–154. doi:10.1016/S0959-4388(02)00308-2

Rizzolatti, G., and Luppino, G. (2001). The cortical motor system. Neuron 31,889–901. doi: 10.1016/S0896-6273(01)00423-8

Rüschemeyer, S.-A., Brass, M., and Friederici, A. D. (2007). Comprehending pre-hending: neural correlates of processing verbs with motor stems. J. Cogn.Neurosci. 19, 855–865. doi: 10.1162/jocn.2007.19.5.855

Schubotz, R. I., and von Cramon, D. Y. (2003). Functional-anatomical conceptsof human premotor cortex: evidence from fMRI and PET studies. Neuroimage20(Suppl. 1), S120–S131. doi: 10.1016/j.neuroimage.2003.09.014

Simon, S. R., Meunier, M., Piettre, L., Berardi, A. M., Segebarth, C. M., andBoussaoud, D. (2002). Spatial attention and memory versus motor preparation:

Frontiers in Human Neuroscience www.frontiersin.org June 2014 | Volume 8 | Article 360 | 9

Moody-Triantis et al. Hand specificity in language

premotor cortex involvement as revealed by fMRI. J. Neurophysiol. 88,2047–2057. Available online at: http://jn.physiology.org/content/88/4/2047.long

Swallow, M. K., Braver, T. S., Snyder, A. Z., Speer, N. K., and Zacks, J. M. (2003).Reliability of functional localization using fMRI. Neuroimage 20, 1561–1577.doi: 10.1016/S1053-8119(03)00436-1

Tanji, J., Okano, K., and Sato, K. C. (1988). Neuronal activity in cortical motor areasrelated to ipsilateral, contralateral, and bilateral digit movements of the monkey.J. Neurophysiol. 60, 325–343.

Tettamanti, M., Buccino, G., Saccuman, M. C., Gallese, V., Danna, M., Scifo, P., et al.(2005). Listening to action-related sentences activates fronto-parietal motorcircuits. J. Cogn. Neurosci. 17, 273–281. doi: 10.1162/0898929053124965

Tkach, D., Reimer, J., and Hatsopoulos, N. G. (2007). Congruent activity duringaction and action observation in motor cortex. J. Neurosci. 27, 13241–13250.doi: 10.1523/JNEUROSCI.2895-07.2007

Tyler, L. K., and Marslen-Wilson, W. (2008). Fronto-temporal brain systems sup-porting spoken language comprehension. Philos. Trans. R. Soc. Lond. B Biol. Sci.363, 1037–1054. doi: 10.1098/rstb.2007.2158

Willems, R. M., Hagoort, P., and Casasanto, D. (2010). Body-specific representa-tions of action verbs: neural evidence from right- and left-handers. Psychol. Sci.21, 67–74. doi: 10.1177/0956797609354072

Woolrich, M. W., Behrens, T. E. J., Beckmann, C. F., Jenkinson, M., and Smith, S.M. (2004). Multi-level linear modelling for FMRI group analysis using Bayesianinference. Neuroimage 21, 1732–1747. doi: 10.1016/j.neuroimage.2003.12.023

Woolrich, M. W., Ripley, B. D., Brady, J. M., and Smith, S. M. T. (2001). TemporalAutocorrelation in Univariate Linear Modelling of FMRI Data. Neuroimage 14,1370–1386. doi: 10.1006/nimg.2001.0931

Worsley, K. J., Evans, A. C., Marrett, S., and Neelin, P. (1992). A three-dimensionalstatistical analysis for CBF activation studies in human brain. J. Cereb. BloodFlow Metab. 42, 900–918.

Conflict of Interest Statement: The authors declare that the research was con-ducted in the absence of any commercial or financial relationships that could beconstrued as a potential conflict of interest.

Received: 09 February 2014; accepted: 11 May 2014; published online: 03 June 2014.Citation: Moody-Triantis C, Humphreys GF and Gennari SP (2014) Hand spe-cific representations in language comprehension. Front. Hum. Neurosci. 8:360. doi:10.3389/fnhum.2014.00360This article was submitted to the journal Frontiers in Human Neuroscience.Copyright © 2014 Moody-Triantis, Humphreys and Gennari. This is an open-access article distributed under the terms of the Creative Commons AttributionLicense (CC BY). The use, distribution or reproduction in other forums is permit-ted, provided the original author(s) or licensor are credited and that the originalpublication in this journal is cited, in accordance with accepted academic practice.No use, distribution or reproduction is permitted which does not comply with theseterms.

Frontiers in Human Neuroscience www.frontiersin.org June 2014 | Volume 8 | Article 360 | 10